The Multi-Agent Learning Seminar is a forum for discussing research related to multi-agent learning, including (deep) multi-agent reinforcement learning (MARL), game theory, emergent complexity and curriculum generation, human-AI interaction (including via natural language), and multi-agent communication. We are a collaborative group with participants from many institutions, including Berkeley, Stanford, Google Brain, OpenAI, Facebook AI Research (FAIR), and other universities. Past speakers have included Michael Littman, Chris Amato, Jacob Andreas, Peter Stone, Thore Graepel, Jeff Clune, Ken Stanley, Dorsa Sadigh, Jakob Foerster, Deepak Pathak, Noam Brown, Aleksandra Faust, and many others. For more information and to request to join, please see: https://sites.google.com/view/berkeleymarl/home

Naveen Raman: Global Rewards in Restless Multi-Armed Bandits
56:09
Naveen Raman: Global Rewards in Restless Multi-Armed Bandits
Siddharth Nayak: LLaMAR: Long-Horizon Planning for Multi-Agent Robots in Partially Observable Envs
42:39
Siddharth Nayak: LLaMAR: Long-Horizon Planning for Multi-Agent Robots in Partially Observable Envs
Siqi Liu: NfgTransformer Equivariant Representation Learning of Normal Form Games
47:30
Siqi Liu: NfgTransformer Equivariant Representation Learning of Normal Form Games
Xueguang Lyu: Understanding Centralized Critics
37:38
Xueguang Lyu: Understanding Centralized Critics
Marc Lanctot: Evaluating Agents using Social Choice Theory
57:59
Marc Lanctot: Evaluating Agents using Social Choice Theory
InstaDeep MARL Team: Growing the MARL software ecosystem in JAX
46:02
InstaDeep MARL Team: Growing the MARL software ecosystem in JAX
Ian Gemp: Approximating Nash Equilibria in Normal Form Games via Stochastic Optimization
54:44
Ian Gemp: Approximating Nash Equilibria in Normal Form Games via Stochastic Optimization
Yi Wu: Diversity-Driven Reinforcement Learning
47:42
Yi Wu: Diversity-Driven Reinforcement Learning
Tom Zahavy: Diversifying AI Towards Creative Chess with AlphaZero
54:02
Tom Zahavy: Diversifying AI Towards Creative Chess with AlphaZero
Keegan Harris - Online Meta-Learning in Games and Bandits
51:33
Keegan Harris - Online Meta-Learning in Games and Bandits
Jessy Lin and Nicholas Tomlin: Decision Oriented Dialogue for Human–AI Collaboration
51:31
Jessy Lin and Nicholas Tomlin: Decision Oriented Dialogue for Human–AI Collaboration
Wolfram Barfuss: Intrinsic fluctuations in reinforcement learning promote cooperation
46:25
Wolfram Barfuss: Intrinsic fluctuations in reinforcement learning promote cooperation
Max Smith: Strategic Knowledge Transfer
53:09
Max Smith: Strategic Knowledge Transfer
Vojtěch Kovařík: Game Theory with Simulation of Other Players
55:59
Vojtěch Kovařík: Game Theory with Simulation of Other Players
Niklas Lauffer: Who Needs to Know - Minimal Knowledge for Optimal Coordination
57:21
Niklas Lauffer: Who Needs to Know - Minimal Knowledge for Optimal Coordination
Sriram Ganapathi: Accelerating Training in Multi Agent RL Through Action Advising
54:51
Sriram Ganapathi: Accelerating Training in Multi Agent RL Through Action Advising
Yali Du: Reinforcement Learning with Human Values
52:40
Yali Du: Reinforcement Learning with Human Values
Robert Lange: Discovering Black Box Optimizers via Evolutionary Meta Learning
56:05
Robert Lange: Discovering Black Box Optimizers via Evolutionary Meta Learning
Zun Li: Combining Tree Search and Population Based Training in Game Theoretic RL
56:59
Zun Li: Combining Tree Search and Population Based Training in Game Theoretic RL
Paul Barde: A Model Based Solution to the Offline MARL Coordination Problem
54:52
Paul Barde: A Model Based Solution to the Offline MARL Coordination Problem
Andrei Lupu and Brandon Cui: Adversarial Diversity in Multi Agent Coordination
51:56
Andrei Lupu and Brandon Cui: Adversarial Diversity in Multi Agent Coordination
Lily Xu: Minimax Regret Optimization Robust Planning for Conservation and Maternal Health
50:45
Lily Xu: Minimax Regret Optimization Robust Planning for Conservation and Maternal Health
Joon Park: Generative Agents: Interactive Simulacra of Human Behavior
50:36
Joon Park: Generative Agents: Interactive Simulacra of Human Behavior
Quentin Bertrand: On the Limitations of the Elo, Real World Games are Transitive, not Additive
52:44
Quentin Bertrand: On the Limitations of the Elo, Real World Games are Transitive, not Additive
Shayegan Omidshafiei: Multiagent Behavioral Analysis
51:35
Shayegan Omidshafiei: Multiagent Behavioral Analysis
Athul Paul Jacob: Learning Effective Policies for for Human-Centric Multi-Agent Systems
55:09
Athul Paul Jacob: Learning Effective Policies for for Human-Centric Multi-Agent Systems
Dianbo Liu and Vedant Shah: Stateful active facilitator
49:24
Dianbo Liu and Vedant Shah: Stateful active facilitator
Miguel Suau: Scaling up MARL: Distributed Simulation of Large Networked Systems
52:10
Miguel Suau: Scaling up MARL: Distributed Simulation of Large Networked Systems
Rose Wang: In the ZONE: Measuring difficulty and progression in curriculum generation
53:19
Rose Wang: In the ZONE: Measuring difficulty and progression in curriculum generation
Frans Oliehoek: Influence-based Abstraction for Faster (Multiagent) Planning and Learning
51:18
Frans Oliehoek: Influence-based Abstraction for Faster (Multiagent) Planning and Learning